Wednesday, May 18, 2016

Trended Credit Data - Pitfalls of Deeper Data Analysis in DU 10.0

As Fannie Mae rolls out DU 10.0, they point out the benefits included in the June 25, 2016 updates, such as the simplification of self-employment categorization and identification of multiple financed properties. But the introduction of trended credit data gets explained as an "improved analysis of risk."  In origination, the phase "improvement" normally means a curtailment of guidelines to the detriment of flexibility, and this update seems to confirm that.

Trended data simply means that the underwriter, live or automated, can evaluate the historical balance and payment amounts on revolving credit. Fannie Mae's position on the face of this is "we can see how much a borrower pays (beyond the minimum payment) to give additional risk flexibility to the borrower who pays more than the minimum regularly."  What Fannie Mae doesn't say, and what presenters don't seem to want to address, is the OTHER information revealed by this data, and its implications for approval and the processing of the application.

For instance, a borrower who pays off credit cards in anticipation of getting a mortgage in the months approaching a home purchase may have the source of those payoff funds questioned, throwing the whole 60 - 90 day seasoning of funds assumption into question. In addition, the underwriter may see a historically higher level of debt, and due to Ability to Repay concerns, apply that level of debt to current qualifying ratios reducing the amount the borrower can afford.

Certainly, a deeper look at a borrower's spending patterns gives a better idea of the risk profile, but opacity of guidelines for reviewing the other aspects and implications of this information means loan production will have more uncertainty - not less, as Fannie Mae posits - regarding underwriting outcomes.